The announcement of Project Stargate comes at a pivotal moment in US tech policy. While OpenAI and SoftBank lead this massive $500 billion private sector investment, the US government is charting its own course through significant policy changes.
The consortium represents an unprecedented collaboration of tech giants:
Core Leadership:
- OpenAI and SoftBank (40% stake each)
- Combined technical innovation and financial power
Strategic Partners:
- Microsoft, Nvidia, and Arm (hardware expertise)
- Oracle (cloud infrastructure)
- MGX (Middle Eastern strategic investment)
Phase one begins in Texas with a million-square-foot facility, laying the groundwork for nationwide expansion.
US Policy Evolution
Just days after Stargate's announcement, President Trump signed an executive order that marks a dramatic shift in US AI policy:
Key Policy Changes:
- Revocation of Biden's 2023 AI safety regulations
- Development of a new AI action plan within 180 days
- Focus on "ideology-free" AI development
The Pentagon has already signalled interest in leveraging Stargate's infrastructure, highlighting the growing convergence of private and military AI development.
The Global Context
The US is of course not the only player in AI:
China's Strategic Vision
China's Next Generation AI Development Plan, launched in 2017, focuses on
- Matching global capabilities
- Achieving major AI breakthroughs by 2025
- Becoming a global AI innovation center by 2030
The recent breakthroughs of Chinese AI labs like DeepSeek and Minimax show they are on track to become an AI powerhouse in the coming years.
EU's Regulatory Leadership
The EU pursues a distinct path focused on governance and ethics:
- AI Act implementation begins February 2025
- Comprehensive regulatory framework through 2026
- Focus on human-centric AI development and innovation
Without clear economic incentives to foster AI innovation the EU will in all likelihood primarily remain a market for AI products rather than a centre of AI developments in the near future.
India's Technology Transformation
India's AI strategy combines ambitious economic goals with social development:
- Projected AI market of $17-22 billion by 2027
- Potential to generate $1.5 trillion in economic value by 2030
- Focus on public-private partnerships and digital infrastructure
What This Means For Your Business
- Regulatory Arbitrage: Different rules mean different opportunities. European operations might need more compliance investment, while India could offer more experimental freedom.
- Market Access: China’s strict data laws might make it harder to deploy global AI solutions there. Consider building region-specific versions of your AI tools.
- Strategic Planning: Watch the US closely – Trump’s deregulation sweep could open new opportunities for rapid AI deployment.
- Risk Management: Don’t put all your AI eggs in one basket. Consider a diversified approach across regions.
The Bottom Line
We’re witnessing the biggest technological gold rush since the internet. While China and India are building AI highways, Europe is installing guard rails, and the US might be about to remove some speed limits.
Smart business leaders will watch these policies closely, adapting their AI strategies to leverage each region’s unique approach. The winners won’t be those who pick sides, but those who learn to play all boards simultaneously.
The Lost Art of Customer Research: Why AI Can’t Save Bad Product-Market Fit
Remember Quibi? The short-form video platform raised $1.75 billion in funding but shut down just six months after launch. They forgot the single most important aspect of marketing: to appeal to their customers.
In today’s AI-obsessed world, we’re seeing history repeat itself. Founders and marketers are so focused on hyping AI that they’re forgetting the basics: deeply understanding their customers and the broader market environment they operate in.
The Niche Trap
While finding your niche is critical (and AI tools can help identify micro-segments), treating niches as isolated islands is dangerous. Even niche products need broader market appeal and ecosystem understanding.
What We’re Forgetting
In the age of mass media, companies had no choice but to understand broad market dynamics. Today’s hyper-targeting capabilities have made us lazy. We think finding a niche is enough, but we’re missing crucial context:
- Macroeconomic factors affecting purchasing power
- Cross-segment influences and trends
- Cultural shifts and societal changes
- Broader industry ecosystem dynamics
The AI Paradox
Ironically, even as AI makes customer research easier, many companies are doing less of it. They’re using AI to optimise messaging and targeting without first validating if their fundamental value proposition makes sense.
The Way Forward
Here’s how to combine AI capabilities with proper customer research:
- Start broad, then narrow: Use AI tools to analyze macro trends first, then drill down to niches
- Connect the dots: Look for overlapping patterns across different customer segments
- Validate assumptions: Use AI for initial insights, but verify with real customer interactions
- Monitor the ecosystem: Track broader market changes that could impact your niche
Have A Gardening Mindset
Niches are seeds, not final destinations. Without understanding the soil (market conditions) and climate (macroeconomic environment), even the best seeds won’t grow. As you embrace AI tools for marketing and customer insights, don’t forget the fundamental art of customer research.
Remember: Quibi didn’t fail because it lacked sophisticated technology. It failed because it misunderstood fundamental human behaviour and market dynamics. Don’t let your AI tools make you forget this crucial lesson.
This week in AI
- OpenAI and SoftBank announced "Project Stargate," a $500B infrastructure initiative over four years, with $100B immediate deployment. The project aims to build massive AI computing infrastructure across the US, starting in Texas. Key equity partners include OpenAI, SoftBank (with Masayoshi Son as chairman), Oracle, and MGX. Technology partnerships include Microsoft, Nvidia, Arm, and Oracle. The stated goals are to secure US AI leadership, create hundreds of thousands of jobs, and support national security interests.
- Chinese AI firm DeepSeek released R1, a breakthrough reasoning-focused language model trained primarily through reinforcement learning. The model achieves performance comparable to OpenAI's O1 on math, code and reasoning tasks at a much lower price pont. Notable features include 671B parameters (37B activated), 128K context window, and pure RL training without traditional supervised fine-tuning.
- Perplexity launched its AI Assistant for Android, marking a shift from search engine to integrated mobile assistant. Features include multi-app actions (ride-hailing, restaurant bookings), camera-based queries, and contextual task handlin. Available in 15 languages including English, Spanish, French, German, Japanese, Korean, and Hindi. Free for all users, though early reviews note some reliability issues that the company plans to address. The launch follows recent Sonar API release and Read.cv acquisition, with Perplexity now valued at $9 billion after raising $500M.
- OpenAI released Operator, an AI agent that can autonomously browse the web and perform tasks using its own browser. Powered by a new Computer-Using Agent (CUA) model combining GPT-4o's vision with reinforcement learning. Initially available only to US Pro users as a research preview. Partners include DoorDash, Instacart, OpenTable, and others. Features built-in safety measures like takeover mode for sensitive information and monitoring for malicious websites. API access to the CUA model planned for developers to build their own agents
